Despite my protestations to the contrary, and my sincere belief that everything was better when we all used command lines, time marches on and my shouting at kids to get off my lawn routine just is not carrying much water any more. We humans are visual creatures, and visualizing the good and the bad helps us make sense of our environment. It is one reason among many why video conferencing has taken off over the last few years; we get so much more out of a conversation when we can see the other person’s facial expressions along with hearing their voice.
So it goes in the world of network troubleshooting as well. For years we stared at text data streaming across our screens, whether from real-time monitoring applications (home grown or otherwise,) or the output of tools we manually ran like ICMP echoes, path traces, etc. But we were limited in what we could see in the proverbial code. We looked at the matrix, but never experienced it.
Eventually we started to take what we already had, our formerly static network maps, and automate them somewhat to show real-time data. Now our NOC operators, or anyone else who regularly looked at the network, could “see” problems as they were happening. And if a picture is worth a thousand words, then a network map is worth a hell of a lot of static textual printouts. But even though we could now see trouble spots in the network, and bathe the boss in the radiating confidence from a big green button, we still had our feet firmly planted in the old ways.
Half-in, half-out might be the best way to view where we had gotten to by the time our green buttons and network maps became firmly lodged in the standard operating practices of NOC operators everywhere. Graphical maps with flashing buttons, and all the real meat came in the form of text boxes filled with data. Even more hampering was the fact that a lot of our visibility came from legacy tools that had not kept up with the ever-changing reality of our networks. Networks which used to be somewhat flat, open, and largely not prone to asymmetrical routing or very complex application stacks. And god knows the cloud was not even a buzzword at that point, let alone a concept with real solutions behind it.
Fast-forward to today where we do have the application stacks, mass virtualization, the cloud, containers, virtualized networks, and a roughly eighty-percent chance of asymmetrical routing both inside and outside of our networks. It is pretty clear that the legacy tools we have been using are just not cut out for today’s reality. What is really needed is a new set of tools, a new way of looking at our networks and our traffic patterns, which take into account all of these challenges we face in modern networks. We need something to keep up with the speed and agility with which our networks and our response times, and the business, demand.
Our tool sets going forward are certainly going to have to cope with a much more complex and ever changing landscape than at any prior time in our history. We’ll need to be able to not just monitor network paths and look for latency which may or may not be artificial, but also to look much more accurately into the performance of our businesses’ applications quickly, visually, and in a way that allows us to hone in on a particular area or device in our networks. A tool to solve all of these complex problems—to help those of us in the network world look like heroes to our bosses—would be a tool in high demand in the marketplace. One step closer to hearing and seeing no evil, and to not having to speak it as well. Of course, we will still give our boss that shiny green button.